maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity

AMDGPU/GlobalISel: Add a few tests for unaligned truncating stores

This commit is contained in:
Matt Arsenault 2021-07-27 10:57:29 -04:00 committed by Matt Arsenault
parent 835ef6f93d
commit 74c65906bc
1 changed files with 146 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

149
llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-store.mir
View File

@ -189,15 +189,58 @@ body: |
; SI: [[COPY:%[0-9]+]]:_(p1) = COPY $vgpr0_vgpr1
; SI: [[COPY1:%[0-9]+]]:_(s64) = COPY $vgpr2_vgpr3
; SI: [[TRUNC:%[0-9]+]]:_(s32) = G_TRUNC [[COPY1]](s64)
; SI: G_STORE [[TRUNC]](s32), [[COPY]](p1) :: (store (s8), addrspace 1)
; SI: G_STORE [[TRUNC]](s32), [[COPY]](p1) :: (store (s16), addrspace 1)
; VI-LABEL: name: test_truncstore_global_s64_to_s16
; VI: [[COPY:%[0-9]+]]:_(p1) = COPY $vgpr0_vgpr1
; VI: [[COPY1:%[0-9]+]]:_(s64) = COPY $vgpr2_vgpr3
; VI: [[TRUNC:%[0-9]+]]:_(s32) = G_TRUNC [[COPY1]](s64)
; VI: G_STORE [[TRUNC]](s32), [[COPY]](p1) :: (store (s8), addrspace 1)
; VI: G_STORE [[TRUNC]](s32), [[COPY]](p1) :: (store (s16), addrspace 1)
%0:_(p1) = COPY $vgpr0_vgpr1
%1:_(s64) = COPY $vgpr2_vgpr3
G_STORE %1, %0 :: (store (s8), addrspace 1)
G_STORE %1, %0 :: (store (s16), addrspace 1)
...
---
name: test_truncstore_global_s64_to_s16_align1
body: |
bb.0:
liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
; SI-LABEL: name: test_truncstore_global_s64_to_s16_align1
; SI: [[COPY:%[0-9]+]]:_(p1) = COPY $vgpr0_vgpr1
; SI: [[COPY1:%[0-9]+]]:_(s64) = COPY $vgpr2_vgpr3
; SI: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s64)
; SI: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; SI: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[UV]], [[C]](s32)
; SI: [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 8
; SI: [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 65535
; SI: [[COPY2:%[0-9]+]]:_(s32) = COPY [[UV]](s32)
; SI: [[AND:%[0-9]+]]:_(s32) = G_AND [[COPY2]], [[C2]]
; SI: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[AND]], [[C1]](s32)
; SI: [[COPY3:%[0-9]+]]:_(s32) = COPY [[UV]](s32)
; SI: G_STORE [[COPY3]](s32), [[COPY]](p1) :: (store (s8), addrspace 1)
; SI: [[C3:%[0-9]+]]:_(s64) = G_CONSTANT i64 1
; SI: [[PTR_ADD:%[0-9]+]]:_(p1) = G_PTR_ADD [[COPY]], [[C3]](s64)
; SI: [[COPY4:%[0-9]+]]:_(s32) = COPY [[LSHR1]](s32)
; SI: G_STORE [[COPY4]](s32), [[PTR_ADD]](p1) :: (store (s8) into unknown-address + 1, addrspace 1)
; VI-LABEL: name: test_truncstore_global_s64_to_s16_align1
; VI: [[COPY:%[0-9]+]]:_(p1) = COPY $vgpr0_vgpr1
; VI: [[COPY1:%[0-9]+]]:_(s64) = COPY $vgpr2_vgpr3
; VI: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s64)
; VI: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[UV]](s32)
; VI: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; VI: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[UV]], [[C]](s32)
; VI: [[C1:%[0-9]+]]:_(s16) = G_CONSTANT i16 8
; VI: [[LSHR1:%[0-9]+]]:_(s16) = G_LSHR [[TRUNC]], [[C1]](s16)
; VI: [[COPY2:%[0-9]+]]:_(s32) = COPY [[UV]](s32)
; VI: G_STORE [[COPY2]](s32), [[COPY]](p1) :: (store (s8), addrspace 1)
; VI: [[C2:%[0-9]+]]:_(s64) = G_CONSTANT i64 1
; VI: [[PTR_ADD:%[0-9]+]]:_(p1) = G_PTR_ADD [[COPY]], [[C2]](s64)
; VI: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[LSHR1]](s16)
; VI: G_STORE [[ANYEXT]](s32), [[PTR_ADD]](p1) :: (store (s8) into unknown-address + 1, addrspace 1)
%0:_(p1) = COPY $vgpr0_vgpr1
%1:_(s64) = COPY $vgpr2_vgpr3
G_STORE %1, %0 :: (store (s16), addrspace 1, align 1)
...
---
@ -221,6 +264,106 @@ body: |
G_STORE %1, %0 :: (store (s32), addrspace 1)
...
---
name: test_truncstore_global_s64_to_s32_align2
body: |
bb.0:
liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
; SI-LABEL: name: test_truncstore_global_s64_to_s32_align2
; SI: [[COPY:%[0-9]+]]:_(p1) = COPY $vgpr0_vgpr1
; SI: [[COPY1:%[0-9]+]]:_(s64) = COPY $vgpr2_vgpr3
; SI: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s64)
; SI: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; SI: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[UV]], [[C]](s32)
; SI: [[COPY2:%[0-9]+]]:_(s32) = COPY [[UV]](s32)
; SI: G_STORE [[COPY2]](s32), [[COPY]](p1) :: (store (s16), addrspace 1)
; SI: [[C1:%[0-9]+]]:_(s64) = G_CONSTANT i64 2
; SI: [[PTR_ADD:%[0-9]+]]:_(p1) = G_PTR_ADD [[COPY]], [[C1]](s64)
; SI: [[COPY3:%[0-9]+]]:_(s32) = COPY [[LSHR]](s32)
; SI: G_STORE [[COPY3]](s32), [[PTR_ADD]](p1) :: (store (s16) into unknown-address + 2, addrspace 1)
; VI-LABEL: name: test_truncstore_global_s64_to_s32_align2
; VI: [[COPY:%[0-9]+]]:_(p1) = COPY $vgpr0_vgpr1
; VI: [[COPY1:%[0-9]+]]:_(s64) = COPY $vgpr2_vgpr3
; VI: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s64)
; VI: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; VI: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[UV]], [[C]](s32)
; VI: [[COPY2:%[0-9]+]]:_(s32) = COPY [[UV]](s32)
; VI: G_STORE [[COPY2]](s32), [[COPY]](p1) :: (store (s16), addrspace 1)
; VI: [[C1:%[0-9]+]]:_(s64) = G_CONSTANT i64 2
; VI: [[PTR_ADD:%[0-9]+]]:_(p1) = G_PTR_ADD [[COPY]], [[C1]](s64)
; VI: [[COPY3:%[0-9]+]]:_(s32) = COPY [[LSHR]](s32)
; VI: G_STORE [[COPY3]](s32), [[PTR_ADD]](p1) :: (store (s16) into unknown-address + 2, addrspace 1)
%0:_(p1) = COPY $vgpr0_vgpr1
%1:_(s64) = COPY $vgpr2_vgpr3
G_STORE %1, %0 :: (store (s32), addrspace 1, align 2)
...
---
name: test_truncstore_global_s64_to_s32_align1
body: |
bb.0:
liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
; SI-LABEL: name: test_truncstore_global_s64_to_s32_align1
; SI: [[COPY:%[0-9]+]]:_(p1) = COPY $vgpr0_vgpr1
; SI: [[COPY1:%[0-9]+]]:_(s64) = COPY $vgpr2_vgpr3
; SI: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s64)
; SI: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; SI: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[UV]], [[C]](s32)
; SI: [[C1:%[0-9]+]]:_(s32) = G_CONSTANT i32 8
; SI: [[COPY2:%[0-9]+]]:_(s32) = COPY [[C1]](s32)
; SI: [[C2:%[0-9]+]]:_(s32) = G_CONSTANT i32 65535
; SI: [[COPY3:%[0-9]+]]:_(s32) = COPY [[UV]](s32)
; SI: [[AND:%[0-9]+]]:_(s32) = G_AND [[COPY3]], [[C2]]
; SI: [[LSHR1:%[0-9]+]]:_(s32) = G_LSHR [[AND]], [[COPY2]](s32)
; SI: [[COPY4:%[0-9]+]]:_(s32) = COPY [[LSHR]](s32)
; SI: [[AND1:%[0-9]+]]:_(s32) = G_AND [[COPY4]], [[C2]]
; SI: [[LSHR2:%[0-9]+]]:_(s32) = G_LSHR [[AND1]], [[C1]](s32)
; SI: [[COPY5:%[0-9]+]]:_(s32) = COPY [[UV]](s32)
; SI: G_STORE [[COPY5]](s32), [[COPY]](p1) :: (store (s8), addrspace 1)
; SI: [[C3:%[0-9]+]]:_(s64) = G_CONSTANT i64 1
; SI: [[PTR_ADD:%[0-9]+]]:_(p1) = G_PTR_ADD [[COPY]], [[C3]](s64)
; SI: [[COPY6:%[0-9]+]]:_(s32) = COPY [[LSHR1]](s32)
; SI: G_STORE [[COPY6]](s32), [[PTR_ADD]](p1) :: (store (s8) into unknown-address + 1, addrspace 1)
; SI: [[C4:%[0-9]+]]:_(s64) = G_CONSTANT i64 2
; SI: [[PTR_ADD1:%[0-9]+]]:_(p1) = G_PTR_ADD [[COPY]], [[C4]](s64)
; SI: [[COPY7:%[0-9]+]]:_(s32) = COPY [[LSHR]](s32)
; SI: G_STORE [[COPY7]](s32), [[PTR_ADD1]](p1) :: (store (s8) into unknown-address + 2, addrspace 1)
; SI: [[C5:%[0-9]+]]:_(s64) = G_CONSTANT i64 3
; SI: [[PTR_ADD2:%[0-9]+]]:_(p1) = G_PTR_ADD [[COPY]], [[C5]](s64)
; SI: [[COPY8:%[0-9]+]]:_(s32) = COPY [[LSHR2]](s32)
; SI: G_STORE [[COPY8]](s32), [[PTR_ADD2]](p1) :: (store (s8) into unknown-address + 3, addrspace 1)
; VI-LABEL: name: test_truncstore_global_s64_to_s32_align1
; VI: [[COPY:%[0-9]+]]:_(p1) = COPY $vgpr0_vgpr1
; VI: [[COPY1:%[0-9]+]]:_(s64) = COPY $vgpr2_vgpr3
; VI: [[UV:%[0-9]+]]:_(s32), [[UV1:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[COPY1]](s64)
; VI: [[TRUNC:%[0-9]+]]:_(s16) = G_TRUNC [[UV]](s32)
; VI: [[C:%[0-9]+]]:_(s32) = G_CONSTANT i32 16
; VI: [[LSHR:%[0-9]+]]:_(s32) = G_LSHR [[UV]], [[C]](s32)
; VI: [[TRUNC1:%[0-9]+]]:_(s16) = G_TRUNC [[LSHR]](s32)
; VI: [[C1:%[0-9]+]]:_(s16) = G_CONSTANT i16 8
; VI: [[LSHR1:%[0-9]+]]:_(s16) = G_LSHR [[TRUNC]], [[C1]](s16)
; VI: [[LSHR2:%[0-9]+]]:_(s16) = G_LSHR [[TRUNC1]], [[C1]](s16)
; VI: [[COPY2:%[0-9]+]]:_(s32) = COPY [[UV]](s32)
; VI: G_STORE [[COPY2]](s32), [[COPY]](p1) :: (store (s8), addrspace 1)
; VI: [[C2:%[0-9]+]]:_(s64) = G_CONSTANT i64 1
; VI: [[PTR_ADD:%[0-9]+]]:_(p1) = G_PTR_ADD [[COPY]], [[C2]](s64)
; VI: [[ANYEXT:%[0-9]+]]:_(s32) = G_ANYEXT [[LSHR1]](s16)
; VI: G_STORE [[ANYEXT]](s32), [[PTR_ADD]](p1) :: (store (s8) into unknown-address + 1, addrspace 1)
; VI: [[C3:%[0-9]+]]:_(s64) = G_CONSTANT i64 2
; VI: [[PTR_ADD1:%[0-9]+]]:_(p1) = G_PTR_ADD [[COPY]], [[C3]](s64)
; VI: [[COPY3:%[0-9]+]]:_(s32) = COPY [[LSHR]](s32)
; VI: G_STORE [[COPY3]](s32), [[PTR_ADD1]](p1) :: (store (s8) into unknown-address + 2, addrspace 1)
; VI: [[C4:%[0-9]+]]:_(s64) = G_CONSTANT i64 3
; VI: [[PTR_ADD2:%[0-9]+]]:_(p1) = G_PTR_ADD [[COPY]], [[C4]](s64)
; VI: [[ANYEXT1:%[0-9]+]]:_(s32) = G_ANYEXT [[LSHR2]](s16)
; VI: G_STORE [[ANYEXT1]](s32), [[PTR_ADD2]](p1) :: (store (s8) into unknown-address + 3, addrspace 1)
%0:_(p1) = COPY $vgpr0_vgpr1
%1:_(s64) = COPY $vgpr2_vgpr3
G_STORE %1, %0 :: (store (s32), addrspace 1, align 1)
...
---
name: test_truncstore_global_s128_to_s16
body: |